Risk and sentiment
US 10Y-3M Spread
The spread between 10-year and 3-month US Treasury yields. Validated by the New York Fed and academic research as the curve with the strongest recession-forecasting record, with the re-steepening after a long inversion historically the more dangerous phase.
At a glance
US 10Y-3M Spread: latest value and prior change
As of 2026-09-04, US 10Y-3M Spread is 0.87 %p. It is 0.01 %p lower than 0.88 %p on 2026-09-03.
Within the default comparison window of 1,000 observations from 2022-09-06 to 2026-09-04, the latest value is at the tie-adjusted 99th percentile. Available history covers 1990-01-02 to 2026-09-04 with 9,173 observations.
- Observed on
- Default comparison window
- 2022-09-06–2026-09-04
- Observations · observed cadence
- 1,000 · Daily
- Data source
- US Treasury
A high or low percentile does not by itself make the indicator positive, negative, or a buy or sell signal.
Time-series chart
US 10Y-3M Spread
1990-01-02–2026-09-04
Long-history series are stored as real provider observations. Index, FX, VIX, and ratio charts use historical backfill where providers expose it; Korean investor flow and margin-credit feeds expand as stable historical endpoints become available.
Recent observations and calculated changes
The ten latest chart observations. Change subtracts the previous observation; intervals vary with holidays, release schedules and missing data. A difference in rates or ratios is not an investment return.
| Observed | Value | Previous date | Difference (current − previous) |
|---|---|---|---|
| 2026-09-04 | 0.87 %p | 2026-09-03 | -0.01 %p |
| 2026-09-03 | 0.88 %p | 2026-09-02 | 0.01 %p |
| 2026-09-02 | 0.87 %p | 2026-09-01 | 0.00 %p |
| 2026-09-01 | 0.87 %p | 2026-08-31 | 0.03 %p |
| 2026-08-31 | 0.84 %p | 2026-08-28 | 0.01 %p |
| 2026-08-28 | 0.83 %p | 2026-08-27 | 0.00 %p |
| 2026-08-27 | 0.83 %p | 2026-08-26 | 0.02 %p |
| 2026-08-26 | 0.81 %p | 2026-08-25 | 0.03 %p |
| 2026-08-25 | 0.78 %p | 2026-08-24 | -0.05 %p |
| 2026-08-24 | 0.83 %p | 2026-08-21 | -0.03 %p |
Source: US Treasury
Interpretation guide
The 10y-3m spread: the raw input behind the Fed's recession odds
The US 10Y-3M spread subtracts the 3-month Treasury rate from the 10-year yield, computed from the Treasury's daily yield curve. It is the maturity pair with the deepest academic validation as a recession predictor — the New York Fed's recession-probability model runs on this spread alone — and because the 3-month rate hugs the current policy rate, it pits today's policy stance against the long-run outlook more cleanly than any other curve.
Why academics chose this curve
Both legs of the 10y-2y curve embed policy expectations, but the 3-month bill is effectively pinned to the rate the Fed sets today. That makes this spread a direct confrontation between the market's future and the Fed's present, and it is the combination with the longest empirical track record in recession forecasting.
- The New York Fed's model converts this single spread into a probability of recession within twelve months.
- With the short leg anchored to policy, most spread moves can be attributed to the 10-year side.
- When the 10Y-2Y inverts first while this curve stays positive, markets are pricing recession as 'eventually', not 'now'.
Inversion, lag, and the re-steepening
Inversions of this curve have preceded past US recessions repeatedly, with a lag averaging around a year but varying widely case by case. Historically the more dangerous phase was not the inversion itself but the moment it unwinds and the curve snaps steeper.
- An inversion lasting quarters rather than weeks carries far more signal weight.
- Re-steepening led by collapsing short rates often coincides with imminent cuts — that is, an approaching downturn — so treat it with caution, not relief.
- Reading the end of an inversion as the end of risk has been the most expensive misread of this indicator.
A checking order for Korean investors
A US recession signal reaches Korea through exports and foreign flows, so moves in this spread call for validation against labor and credit data. The standard practice is to watch whether the warning spreads to other markets rather than conclude from the curve alone.
- Overlay it with the US 10Y-2Y Spread on TapeFlow; the sequence of inversions and un-inversions across the two curves refines the read.
- Cross-validate with US Initial Jobless Claims and US High Yield OAS deteriorating in the same direction.
- Use an inversion as a prompt to audit portfolio defenses during the window, not as a sell trigger by itself.